BACKGROUND: It is unclear whether the serum prolactin (PRL) surge following electroconvulsive therapy (ECT) is a marker of optimal ECT administration. We investigated the relations among PRL surge, stimulus parameters, and outcome in major depressive disorder (MDD). METHODS: Seventy-nine patients with MDD were randomized in a double-blind trial to right unilateral (RUL) or bilateral (BL), and to low-dose (just above seizure threshold) or high-dose (2.5 x threshold) ECT. RESULTS: Change in PRL (delta PRL) varied among treatment groups, with significant effects of electrode placement (BL > RUL, p < .006), electrical dosage (high > low, p < .04), and gender (female > male, p < .005). There was no evidence that clinical improvement was associated with greater PRL surge. CONCLUSIONS: Although delta PRL varied with parameters impacting on response rates, these data indicate the PRL surge cannot serve as a useful index of clinically effective treatment. This finding does not support the view that diencephalic seizure propagation is necessary for ECT to exert therapeutic effects.